of of

Memorial University

Investigation of the Genetic Cause of Hearing Loss in 28 Autosomal Dominant Families within the Newfoundland Founder Population

Masters of Science (Medicine)

David A. McComiskey September 2010

Abstract

The purpose of this study was to detennine the genetic cause of hearing loss in28 Newfoundland families with Autosomal Dominant hearing loss. AD hearing loss is highly genetically heterogeneous, and is mainly associated with a lateonset,progressive phenotype. After a comprehensive literature search,genotype-phenotypeevaluations,and afunctionalcandidategeneapproach,all28probandsweresequencedtoidentify mutations in four genes known to cause autosomal dominant hearing loss, COCH, KCNQ4, TECTA, and MYOJA. First, a known Dutch founder mutation within exon 4 of COCH, c.151 C>CT, was found in a Newfoundland proband of Family 2094. All affected family members (n=7) shared this mutation, while unaffected members did not. This is only the second family found to harbor this mutation outside of Europe. This mutation is strongly associated with severe vestibular decline. Affected Family 2094 members carrying the mutation do present vestibular decline in the fonn of vertigo and balance difficulties. As this mutation is considered to be a Dutch founder mutation, DNA samples from a Dutch p.P51P/S family were genotyped and compared with Family 2094 genotypes. Fragment analysis confinned haplotype sharing of five markers closely bordering the c.151 C>CT mutation between Newfoundland and Dutch mutation carriers.

Second, a novel3bp deletion in exon 5 of KCNQ4 was found in 13 affected members of Family 2071. While the mutation was not seen in four other affected family members, audiology test results suggest that these four individuals arephenocopies.Sequencingof the full KCNQ4 gene was done in all individuals, to rule out another mutation on the same gene. Further investigation, through the construction of an intragenic haplotype,did

notpointtoanyfurtherhearinglossassociatedvariantswithinKCNQ4,andconfirmed that all deletion carriers share a common hearing loss haplotype and deletion. Third,a nonsense mutation was found in exon 4 of MYOIA in the proband of Newfoundland Family 2102. This is a C7T nucleotide substitution (c.2435 C>CT) that causes a change (p.R93X) in the motor domain of myosin lAo Offour individuals in Family 2102, three were found to carry the p.R93X mutation, while one unaffected sibling was not. This mutation has been reported once before in a small Italian family. No mutations were discovered in the TECTA gene. When each of the causative mutations in COCH, KCNQ4, and MYOIA was detected, additional Newfoundland hearing loss probands were screened, to rule out the possibilityofa founder mutation.Inno case were additional mutation carriers identified. While no founder mutations were discovered in this study, the genetic cause of hearing loss was identified in three families.

Acknowledgments

I thoroughly enjoyed this graduate program. It was full of rewarding experiencesandI am sad to see it end. I would like to take this opportunity to thank those who aided me and without whom I likely never would have succeeded. First, I would like to thank the lab staff, includingMr.Dante Galutria,Mr.Jim Houston, and Ms. Annette Greenslade, for their patient and unwavering encouragement. Secondly, I would liketothankmy fellow students in the lab, including Lance Doucette, Nelly Abdelfatah, and Jessica Squires for their optimism, encouragement, and happy dispositions. Third, I thank Carol Negrijn, for her invaluable clinical, medical, and computer knowledge. And finally, I wish to thank my supervisor, Dr. Terry-Lynn Young, and my thesis committee members Dr. Ban Younghusband and Dr. Jane Green for pushing me to push myself, in order to bring out the best inmyself,and the best in my research. The things I've learned and the bonds I've forged during my time in "The Young Lab" will never be forgotten.

Table of Contents

Acknowledgements

List of Figures .

ListofAbbreviationsandSymbols ..

List of Appendices

...i

Table of Contents

Cbapterl:lntroduction 6

... 6

Pedigrees 8

Audiograms . 9

Autosomal Dominant Hearing Loss II

Critical Considerations When Researching Autosomal Dominant Hearing Loss 12 The Pioneering of Hearing Loss Gene Discovery ....

Founder Populations&Mutations . Colonization of Newfoundland: A Founder Population

. 18

... 21

. 22

Cbapter2:Metbods&Materials 31

HumanSubjects . 31

Experimental Design: Functional Candidate Gene Mutation Screening.... . 32 General Strategy for PCR and Sequencing of Candidate Genes 35 DNA Preparation, PCR Therrnocycling, and Electrophoresis . 35

Preparation for ABI Cycle Sequencing.... . 36

Automated Sequencing Using the ABI 3130 . 37

Tracing Variants Through Families: Genotype&Haplotype Analysis 38

Cbapter3:Results.... . 49

... 40

Family 2094.... . 50

Search Fora Vestibular Phenotype in Family 2094 Mutation Carriers 51

IdentificationofaDutchFounderMutation 52

Family 2071.... . .

Family2102 .

Cbapter4:Discussion 80

Family 2094 Hearing Loss Caused by COCHMutation.... . . 80

Confirmation of p.PSIPIS as a Dutch Founder Mutation . Family 2071 Hearing Loss Caused by Novel KCNQ4 Deletion ..

Family 2102 Hearing Loss Caused by MYOIA Mutation 88

Candidate Gene TECTA.... .. 91

Non-Founder Mutationsina Founder Populations ..

The Changing Landscape of Gene Identification Methodology ...

Limitations of this Study ...

Chapter 5: Summary.... .. .

Literature Cited.

.. 94

.. 96

.. 99

List of Tables

Table 1.1 AD Non-Syndromic Deafuess Genes Identified to Date (Van Camp G, Smith RJH.Hereditary Hearing Loss Homepage. http://hereditaryhearingloss.org, May 2010) 29 Table 1.2 Deafness Genes Currently Identified in the Newfoundland Population 30 Table 2.1 Candidate Genes for Newfoundland Families Having Late Onset AD Hearing

Loss (Adapted from Hilgert et al. 2009) 48

Table 3.1 Candidate Genes Screened for Mutations in Newfoundland Families Having

Late Onset Autosomal Dominant Hearing Loss 71

Table 3.2 Audiology Summary for Family 2094 Family Members 72

Table 3.3 Phenotype Summary of Family 2094 Individuals 73

Table 3.4 Physical Location of Markers Used to Create the p.P5IP/S Deafuess Haplotype. The Markers Were Taken From Fransen et al, 2001 74 Table 3.5 Haplotype Sharing Across Markers Flanking the COCH Gene Between

Newfoundland Family 2094 and a Dutch p.P51PISFamily 75

Table 3.6 Phenotype Features of Affected Family 2071 Individuals. Shown First Are Family Members With the Deletion, and Second, Those Without the Deletion . Table 3.7 Audiology Testing Results of Affected Family 2071 Individuals . Table 3.8 KCNQ4 Variants Used to Create the Intragenic KCNQ4 Haplotype Table 3.9 Audiological Summary of Family 2102 Individuals With&Without the p.R93X Nonsense Mutation....

Table 5.1 Deafness Genes Identified in the Newfoundland Population at End of This

Study 102

List of Figures

Figure 1.1 Examples of Audiograms . 26

Figure 1.2 Examples of More Complex Audiograms. . 27

Figure 1.3 Map of the Island of Newfoundland 28

Figure 2.1 28 Autosomal Dominant Newfoundland Pedigrrees 40

Figure 2.2 Flow Chart Demonstrating Experimental Design and Progression 46 Figure 3.1 A Six Generation Newfoundland Family (2094) Segregating an Autosomal Dominant Form of Late Onset Progressive Hearing Loss (partial Pedigree) 56

Figure 3.3 Hearing Loss Phenotype ofCOCHp.P5 IP/S Carriers III-12, V-I, and IV-I Figure 3.4 Genetic Map of Markers Used to Construct the p.P5IP/S Deafness Haplotype

for Newfoundland&Dutch Carriers. .. . 59

Figure 3.5 p.P5IP/S Deafness Haplotype for Newfoundland&Dutch Carriers. . 60

Figure 3.7 A Five Generation Newfoundland Family Affected With Autosomal

Dominant, Late Onset, Progressive Hearing Loss. . 63

Figure 3.8 Electropherogram of the 3 bp Deletion (p.Ser269del) . 64 Figure 3.9 Audiological Summary of Family 2071 Family Members 65

Figure 3.10 Family 2071 Pedigree With Haplotype . 66

Figure 3.11 Structure ofKCNQ4 ...

Figure 3.12 Pedigree of Newfoundland Family 2102 . 68

Figure 3.13 Electropherogram of p.R93X Mutation inMYOlA.... . 69 Figure 3.14 Hearing Loss Phenotype ofMY01ANonsense Mutation Carriers IV-I, III-I,

and IV-5. Onset of Hearing Loss is 5 Years of Age 70

List of Abbreviations&Symbols ABIApplied Biosystems International

AD Autosomal Dominant

ADNSHLAutosomal Dominant Non-Syndromic Hearing Loss

AHCAuditory Hair Cell

C9orf75C90penreadingframe75

eMCentimorgan

COCHCoagulation Factor C Homolog

CPCytoplasmic

CTComputed Tomography

DFNAAutosomal dominant deafness gene

DFNBAutosomal recessive deafness gene

DFNX-Linked deafness gene

dH20De-ionizedwater

List of Abbreviations&Symbols (cont)

DIAPHI Protein diaphanous homolog I dNTP Dideoxynuc1eotide Triphosphate

DMSO Dimethyl Sulfoxide

DNA Deoxyribonuc1eicacid

EDTA Ethylenediaminetetraacetic acid

EYA4 Eyes absent homolog 4

FCH Factor C Homologous Domain

GS500(-250)LIZ GeneScan500(-250)LIZSizeStandard

GJB2 Gapjunctionprotein,beta2

GJB6 Gapjunctionprotein,beta6

GWS Genome Wide Scan mcHuman Investigations Committee

KCNQ4 Potassium Voltage-Gated Channel 4

List of Abbreviations&Symbols (cont)

LDLinkage Disequilibrium

MbMegabase

MgChMagnesium Chloride

MP3MPEG-l Audio Layer 3

MSH2MutShomolog2

MYOIAMyosin IA

MY07AMyosin VIlA

ngNanogram

OMlMOnline Mendelian Inheritance in Man

PAX3Paired box gene

PCRPolymerase Chain Reaction

rpmrevolutionsperminute

List of Abbreviations& Symbols (cont)

SSCPSingle strand confonnation polymorphism

SNPSingle Nucleotide Polymorphism

TaqThennus aquaticus DNA Polymerase

THETris/BoratelEDTA Buffer

TECTATectorin Alpha

TMPRSS3Transmembrane protease 3, serine 3

TORCHToxoplasmosis, rubella, cmv, and herpes

TPRNTaperin

U Units

,Ill Microlitre

,uM Micromolar

USA United States of America

USHUsher Syndrome

WFSl Wolfram syndrome I (wolframin)

WSWaardenburgSyndrome

List of Appendices

Appendix A Mutations previously found within the four selected candidate genes KCNQ4, COCH, TECTA, and MY01A Primer Sequences and Expected PCR Product

SizesOfAlIExonsSequenced... . .... 112

Appendix B COCH Microsatellite Marker Primer Sequences and Expected PCR Product

Size.. .. 115

Appendix C COCHMicrosatellite Marker Genotype Data .. .. .117

Chapter 1: Introduction

Purpose

The aim of this research project is to determine the genetic etiology of autosomal dominant (AD) hearing loss in 28 Newfoundland families.

Overview

Hearing loss is the most common sensory disorder in humans. For example, one in every 500 newborns has hearing loss (Morton&Nance, 2006). The prevalence of hearing loss increases dramaticaJly with age, and by puberty, the number of affected persons doubles (Morton&Nance, 2006). Hearing loss is even more prevalent in adults, as 60%of people older than 70 years have a hearing loss of25 dB or more (Gratton&

Vazquez,2003).

Hearing loss is a multi-factorial disorder caused by both genetic and environmental factors. Genetic factors account for 50 % of all hearing loss cases, while environmental factors cause 25%.The remaining 25%are classified as being of unknown etiology (Willems, 2001). Environmental causes of hearing loss include exposure to high sound decibel levels, head trauma, prematurity,neonatal hypoxia, low birth weight, prenatal infections from "TORCH" organisms (i.e., toxoplasmosis, rubella, CMV, and herpes), and postnatal infections like bacterial meningitis (Willems, 2001;

Bitner-Glindzicz, 2002).

Approximately 30 % of genetic cases aresyndromic: the phenotype includes0ther signs and symptoms throughout the body in addition to deafuess. Over 400 genetic syndromes include some degree of hearing loss (Gorlin et al. 1995; Nie et al. 2008). Two examples are Usher syndrome (USH): hearing loss accompanied by retinitis pigmentosa, and Pendred syndrome: a hearing loss disorder accompanied by goiter, whichisa swelling in the thyroid gland. However, the vast majority, around 70%, of inherited hearing disorders are non-syndromic (Cremers et al. 1991; Van Camp et a!. 1997).

Worldwide, within non-syndromic cases, 88 % of the hearing loss genes identifiedcause autosomal recessive (AR) hearing loss, II % AD, and the remaining 1% either mitochondrial or X-linked (Smith & Van Camp, 2007).

Thefivefactorsusedtodescribehearinglossareageofonset,soundfrequencies affected (low, middle or high), degree of hearing loss (measured indBs), affected part of the auditory system (conductive, sensorineural or mixed), and configuration(unilateral, or bilateral).

Hearing loss has a high degree of genetic heterogeneity. A large numher of mutations within many different genes cause similar hearing loss phenotypes. As of May 2010, the Hereditary Hearing Loss Homepage listed 141 non-syndromic deafness loci that haveheen mapped and 50 genes that have been identified. Twenty-two0fthe50 known genes harbor mutations that cause AD hearing loss (Table 1.1), 33 cause AR hearing loss, and 2 cause X-linked hearing loss (Van Camp G, SmithRJH http://hereditaryhearingloss.org). Loci for non-syndromic hearing loss are denoted 'OFNA' for AD inheritance, 'DFNB' for AR inheritance and 'OFN' for X-linked

inheritance (Griffith&Friedman, 2002). Some genes cause both AD and AR hearing loss. For example, Grifa et al. (1999) found a C7T change in gap junction protein, beta 6 (GJB6) that resulted in the substitution ofa highly conserved threonine residuefora methionine at amino acid position 5 (p.T5M), resulting in nonsyndromic AD hearing loss. While thisGJB26 mutation causes AD hearing loss, Del Castillo et al. (2002) identified a342 kb deletion inGJB6by studying 422 unrelated subjectsfromSpainand Cuba with an AR pattern of inheritance.

Pedigrees

When studying hearing loss, or any hereditary disorder, family members are visualized on a pedigree chart, which in this study shows all known hearing10ss phenotypes presented at the time of clinical and audiological testing.Thisallowseasier identification of the inheritance pattern and of the relationshipsamonghaplotypes.A haplotype is a combination of alleles that are transmitted together. When a causative mutation is found, alleles oflinked markers are assessed in order to develop a haplotype, or pedigree that illustrates shared genetic variants between farnily members.These haplotypes are then compared between members of the same family or between members of different families that share the same mutation. A common haplotype with the same mutation suggests a common ancestor for that mutation. Furthermore, a haplotype can point to associations between different variants that may be combining to affect the phenotype.

Audiograms

Audiograms are graphs of the minimal level of sound that a given person can hear at various frequencies (Figure 1.1; Figure 1.2). They are produced using an audiometer, a machine that tests hearing by exposing patients to a range of sounds at differentpitches and decibel (dB) levels. During hearing tests, separate audiograrns are obtainedforeach ear. Each line on the audiogram represents one ear. The y-axis measures sound intensity in units of dB, which increases logarithmically. The x-axis of the audiogram measures the frequency, or pitch, of a sound in Hz (Hertz). Low pitch sounds have low frequencies« 500 Hz), medium pitch sounds have medium frequencies (500 - 2000 Hz), and high pitch sounds have high frequencies(>2000 Hz). Hearing loss is characterizedbyintensity, which can be mild, moderate,severeorprofound,and by which frequency isaffected, such as low, middle or high.

Anindividual with normal hearing can detect soundsbetweenOdB and 20 dB.

The minimum level of hearing, 0 dB, is equivalent to a barely audible whisper. Those affected with hearing loss, however, have a higher than normal minimum hearing level.

This means that any given sound intensitymustbegreaterthanOdB for them to hear il.

People with a mild degree of hearing loss can only hear sound at intensitiesbetween20- 40 dB for the frequencies of 500 - 4000 Hz. Individuals with moderate hearing loss can only hear sound from40-70dB,andthosewith severe hearing loss can only hear sound between 70-95 dB in intensity. Lastly, those with profound hearing loss cannot detect soundatallunlessitis95dBorgreater(suchasthesoundproducedbyan.MP3player at maximum volume; Mazzoli et aI. 2003).

Figure 1.1 shows a series of simple audiograms: audiogram A shows an individual with normal hearing, B an individual with moderate bilateral (affecting both ears)hearingloss,andCanindividualwithseverebilateralhearingloss.However, audiograms are often not so simple to read. Figure 1.2 shows two additional audiograms:

audiogram A shows an individual with moderate to mild hearing loss in the left ear and normal hearing in the right ear (unilateral),and B shows an individual with bilateral hearing loss sloping to moderate and profound at the higher frequencies. This audioprofile is typical ofpresbyacusis, or age-related hearing loss. 40 % of the population older than 65 yearsofageisaffected,and 80 % of hearing loss cases occur in elderly people (Gates&Mills, 2005).It is now generally accepted that presbyacusis is most often caused by age-related declines in the auditory system, such as lossordeteriorationof sensory cells within the cochlea. Moreover,impairedtemporalprocessingisassociated with age-related factors that affect neural synchrony of hearing (Schuknechtetal.1993;

Friedman 2003; Wu etaI.2003; Fitzgibbons et al. 2010). Temporal processing refers to the processing of acoustic stimuli overtime. Temporal processing allows us to distinguish speech from background noise, as the decibel levels of the background noise

Another common and important characteristic of presbyacusis, and of any sensorineural hearing loss, is the level of speech discrimination a patient demonstrates.

Hearing a sound does not always translate into properly distinguishing speech. Tests are also performed to determine a patient's speech discrimination. The measure of speech discrimination is often a percentage, and describes the ability ofa patient to correctly

identifY words when the sound is loud enough for them to comfortably hear. When a patienthaslowspeechdiscrimination,ahearingaidwillsuccessfullyamplifYsoundin the patient's ear, but will not necessarily improve speech perception. The amplified sound remains gibberish to the patient because he/she is unable to identifY the words.

(McAlister, 1990; Kodera et al. 1994). A cochlear implant, a surgically implanted electronic device that provides sound to profoundly deaf or severely hard of hearing individuals,hasbeenfoundinmanycasestomarkedlyimprovespeechdiscrimination (Leung eta1. 2005; Cambron, 2006; Yueh&Shekelle, 2007).

Autosomal Dominant Hearing Loss

Autosomal dominant non-syndromic hearing loss (ADNSHL) accounts for approximately 15 % of inherited hearing loss (Hildebrand etal. 2008). To date, 59 loci for ADNSHL have been identified, along with 22 causally related genes (Table 1.1; Van Camp G, SmithRJH.Hereditary Hearing Loss Homepage.

hllp://hereditaryhearingloss.org). The majority of AR hearing loss cases are caused by mutations in just a few genes, most notably gap junction protein beta 2(GJB2) and GJB6.

This contrasts sharply with AD hearing loss, which is significantly more genetically heterogeneous (Griffith&Friedman,2002),makingcost-effectivescreening for diagnosis in a clinical selling highly problematic. Mutations within the genes wolfram syndrome I (WFSI), cochlin (COClf), potassium voltage-gated channel 4 (KCNQ:f), and tectorin alpha(TECTA)are marginally more frequently reported in comparison to the

other reported causative genes. The audioprofilesometimesprovides clues to the underlying causative gene. For example,WFSJharbors mutations found in 75 % of families segregating for AD, non-syndromichearing loss that initially affect only the low frequencies(youngetal. 2001; Bespolovaetal. 2001).

ADNSHL is often characterized by a post-lingual, late-onset, progressive phenotype that affects mainly adults. Post-lingual hearing loss is muchmorefrequent thanpre-lingualhearingloss,andaffectsI0%ofthepopulationbytheageof60(Van Camp et al. 1997). This most often results from damage to auditory hair cells (AHCs) or their innervation (Gates&Mills, 2005). For example, one late-onset progressive hearing loss associated gene is eyes absent homolog 4(EYA4),a member of the vertebrateEya family of transcriptional activators. Mutations in this gene were found in Belgian and USA families, and create premature stop codons leading to post-lingual, progressive, AD hearing loss.EYA4was subsequently shown to be critical in the continued function of the mature organ of Corti, an organ in the cocWea that contains the AHCs (Wayne eta1.

2001).

Critical Considerations When Researching Autosomal Dominant Hearing Loss

Of the 28 farnilies researched in this study that are classified as having an AD pattem of inheritance, one or more may have been incorrectly classified duetoalackof sufficient data. Many individuals in these pedigrees (Figure 2.1) areascertained through

relatives' word of mouth. For this reason, itisimportanttodiscusstherolethatdifferent factors may be playing in confusing the ascertainment of individuals and thus the search for causative hearing loss mutations.

Digenic inheritance is when a phenotype is expressed only if an interaction between two mutant alleles in two separate genes occurs (Strachan&Read, 2003).

Digenic inheritance does not cause AD hearing loss, butdigenic inheritance may play a role in the hearing loss of one of the families under investigation in this srudy. For example, Chen et al. (1997) reported a small consanguineous family with three affected and three unaffected members. Two regions shared by the three affected individuals were identified, one on 3q21.3-q25.2 (LOD=2.78) and 19p13.3-p13.1 (LOD=2.78). LOD (Logarithm (base 10) of odds) isa statistical test used to determine the likelihood of obtaining test data if two loci are linked compared to the likelihood ofobserving the data by chance. Chenetal. (1997) speculated that two non-allelic recessivemutations accounted for the profound congenital deafness in this family.Ina Chinese family, Liu et al. (2009) demonstrated through DNA sequencing that mutations in GJB2 and GJB3 interact to cause hearing loss in digenic heterozygotes. To supportthis, they discovered overlapping expression patterns of GJB2 and GJB3 in the cocWea, along with co- assembly of the GJB2 and GJB3 proteins when co-transfectedinhuman embryonic kidney (HEK) cells (Liu et al. 2009). And a third example was seen recently when mutations within ATP sensitive inward rectifier potassium channel 10(KCNJI0)and solute carrier family 26, member 4 (SLC26A4) were found to cause digenic non- syndromichearing loss associated with enlargedvestibularaqueductsyndrome(EVA)

(Yang etaI.2009). Mutations in SLC26A4 were previously shown to cause Pendred syndrome (PS), agenetic disorder leading to hearing loss and goiterwithoccasional hypothyroidism. Many individuals with an EVAlPS phenotype had only one disease- causing variant in SLC26A4. Yang et al. (2009) identified double heterozygosity in affected individuals from two separate families. These patients carry single mutations in both KCNJIO and SLC26A4, and the mutation in SLC26A4 has been previously associated with the EVAlPS phenotype. The KCNJIO mutation reduces potassium conductance activity, which is critical for generating and maintaining proper ion homeostasis in the ear. To add further support to theirdigenic interaction hypothesis, Yang et al. (2009) demonstrated haploinsufficiency of Slc26a4 in Slc26a4+/- mouse mutants resulted in reduced protein expression of KcnjIO in the inner ear.

One important term to keep in mind when researching AD hearing loss is penetrance.Penetrancereferstotheproportionofindividualswitha mutation who exhibit clinical symptoms. For example, if a mutation in a gene responsible for a type of AD hearing loss is 95%penetrant, then 95%of individuals with the mutation will exhibit symptoms, while 5 % will not during their lifetime. Penetrance is often expressed as a frequency at different ages because, for many hereditarydiseases,onsetofsymptoms is age-related (Strachan&Read, 2003). This is particularly important because AD hearing loss is often late-onset and progressive. Forthisreason,afamily'sinheritance pattemcould appear to be sporadic, when in fact the disorder segregatesautosomal dominantly, and the individuals under study simply haven't yet presented the hearing loss phenotype, as the age of onset varies widely and can range well into 50 years of age. A

related but distinct potential problem is expressivity. Expressivityrefers to variations ofa phenotype for a particular genotype. When a condition has higWy variable signsand symptoms, it can be difficult to diagnose.

Mitochondrial inheritance could also be confusing the ascertainment ofthe families investigated in this study. Mitochondrial inheritance is the inheritanceofatrait encodedinthe mitochondrial genome, and is always of maternal origin.It is therefore often also called maternal inheritance. When a woman harbors a mitochondrial mutation, and her egg cells are forming an ovary, these egg cells contain a random distribution of both normal and mutated copies of the mitochondrial gene(St.Johnetal.20 10).

Therefore, all children of this mother may inherit some mutated mitochondria, but if the number of mutated mitochondria reaches a critical level, termed the"thresholdeffect", then an adverse phenotype is seen (St. John et al. 2010; Van Camp G, Smith RJH.

Hereditary Hearing Loss Homepage). These mitochondrial hearing loss mutations can be late-onset, or if the carrier is administered certain antibiotics, the phenotypewillbe

"drawn out". This is the case with mutations inMT-RNRI,which are known to cause non-syndromic deafness (Casano et al. 1999; Bates 2003). Mitochondrial mutations are beyond the scope of this study, but have previously been shown to cause late-onset hearing loss that is comparable to the phenotypes of Families 2093, 2112,and2125 investigated here (Casano et al. 1999). While there are no incidences of male to male transmission in these families, additional clinical ascertainrnentcouldpotentially reveal a mitochondrial inheritance pattern. This study primarily targets only genesandmutations known to be associated with AD hearing loss, but the possibility of maternally inherited

mutations causing hearing loss in the above mentioned families should not be ruled out, and should be investigated in future studies.

As a result of random genetic drift in the founder population of Newfoundland, there is an elevated incidence of particular rare disorders, such as Bardet-Biedlsyndrome (Webb et al. 2009). This makes the founder population of Newfoundland ideal for the study of genetic disorders and increases the chance of detecting novelcausativegenes and mutations. However, due to the nature of Newfoundland as a genetic isolate, some potential pitfalls arise. One potential pitfall is the uncertainty ininheritance ascertainment. For example, assortative mating could confuse the ascertainment of families and therefore the search for hearing loss mutations. Assortative mating occurs when sexually reproducing organisms choose to mate with individuals that are similar (positiveassortativemating)ordissimilar(negativeassortativemating) to themselves in some specific way. One family under investigation in this study (Family 2069) is a potential example of positive assortativemating (seep.42, bottom pedigree,sth generation). This is critically important. Positive assortative matingcouldresultinboth parents carrying a mutation that causes hearing loss. This may, however, simply be a result of studying a genetic disorder in a highly isolated population, and it is possible that thismatingtookplacenotbecausebothindividualswereaffectedbyhearingloss,but instead due simply to the low level of mating choice in small out-port community. Either way, our current inheritance classification of Newfoundland Family 2069 could possibly be incorrectly stated as AD, and our candidate gene selection would thenbe based on unfounded and false assumptions.It is important to keep this possibility under

unfounded and false assumptions. It is important to keep this possibility under consideration, and to investigate and screen for commonly occurring recessivemutations intbe proband of Family 2069 as well as dominant mutations. Pseudo-dominanceshould also be taken intoaccounl. This istbe situation where tbe inheritance ofan ARtrait mimics an AD pattem, and due to tbe limited extent of clinical ascertainment intbese families' histories, it is possible tbat one oftbese AD families is in fact affected by an AR mutation segregating in a pseudo-dominant fashion.

Anotbercriticalfactortojudgewhenresearchinggenetichearinglossistbe possible presence of phenocopies. A phenocopy is an affected individualwhohastbe same disease, but due to a different cause, as relatives affected witb tbegeneticcondition under study. Hearing loss is a very common type of sensory loss in humans. Many types ofenvironmentalandgeneticfactorsaccountforhearinglosssoindividualswitbin families affected witb hearing loss can beaftlicted due to apletboraofdifferentreasons (Griffith & Friedman 2002). For example, a study of heterozygousWFSI mutations in two low frequency sensorineural hearing loss families showed tbat tbese two families' hearing loss were linked to adjacent but non-overlapping loci on4pl6, DFNA6and DFNAI4 (Van Camp et ai. 1999). Upon further study, it was found tbat an individual in tbe DFNA6 family who had a recombination event excluding tbe DFNAl4 candidate region was actually a phenocopy. The cause of hearing loss in this phenocopy was reported as unknown, but as a consequence tbey were able to detennine tbat DFNA6 and DFNAI4areallelic(BespalovaetaI.2001).

Throughout this study genes are investigated through targeted genesequencing.

However, it mustbe mentioned that it is possible larger genomic abnormalities may account for hearing loss in some of the Newfoundland families under investigation (Lisenkaetal. 2003; Shafferetal.2006). Large genomic rearrangements, deletions, inversions, etc. can cause and affect the degree and severity of diseases, and these large- scale anomalies are not detected through traditional DNA sequencing methods (Lisenka etal. 2003; Idbaihetal. 2010).

The Pioneering of Hearing Loss Gene Discovery

The fIrst genes to be implicated in hearing loss were found in the syndcomic disorders. Syndromic forms of hearing loss are classified by their associated symptoms.

For example, Waardenburg syndrome (WS) is the most common cause of AD syndromic hearing loss. WS is characterized by varying degrees of hearing loss associatedwith pigmentation anomalies and neural crest defects.It was frrstdescribedin 1951, but it took several decades to identify the causative genes. Asher and Friedrnan (1990) studied mice and hamsters with four mutations causing phenotypes similar to those seen in human WS patients. They used the chromosomal locations and syntenic relationships associatedwith three of these four mutant mouse genes to predict human chromosomal locations for the causativeWSgene.Syntenyisthesituationwherebyorganismsofrelativelyrecent divergence show similar blocks of genes in the same relative positions in the genome.

Asher and Friedman (1990) predicted four possible locations forthecausative gene, and

one turned out to be correct. In 1992,mutationscausing WS were discoveredinthe paired box gene(PAX3)gene (on chromosome 2q) (Tassabehji et al. 1992). A second common cause of AR syndromic hearing loss is Usher syndrome (Toriello et al. 2004).

Usher syndrome was first described in 1858 when Van Graefe reported the case of a deaf and "dumb" male patient presenting with retinal pigment degeneration, who had two similarly affected brothers (Van Graefe, 1858). This was the first syndrome to demonstrate that phenotypes, in this case deafness and blindness, could be inherited in tandem.Ushersyndromeischaracterizedbyprofoundcongenitalhearingimpairment, retinitispigmentosa, and vestibular dysfunction. It has three clinical types,denotedasI, II,andlII,indecreasingorderofseverity(Saihanetal. 2009). In 1995 one of several causative genes for Usher Syndrome Type I was discovered. WeiI et aJ. (1995) chose myosin 7A(MYOVlIA)as a functional candidate gene, based on observations that cytoskeletal abnormalities seeninUsher syndrome patients are also seen in mouse mutants with myosin mutations. Two different premature stop codons, a six bp deletion, and two different missense mutations were detected in five unrelated families.Inone family, these mutations were identified in both alleles, and resulted in the absence ofa functionalproteinandsubsequentUshersyndrome(WeiletaJ.I995).Currently,IO different types of Usher syndrome have been recognized, with more than 100 pathogenic mutations alone for the two most common molecular forms, Usher IB (USHI B) and Usher 2a (USH2A; Ahmed etal. 2003; SaihanetaJ.2009).

The first ADNSHL family investigated was from the small town of Taras, Costa Rica. The hearing loss was described as low frequency AD with a post-lingual age of

onset at 10 years of age (Leon etaI.1981). Leon et al. (l992)perfonned linkage analysis to detennine that the causative gene was linked to markers defining a 7 cM critical region on chromosome 5q31. Genetic markers are DNA sequences with a known location on a chromosome, and are useful in linkage analysis because they are easily identifiable, associated with a specific locus, and highly polymorphic. LODscoresfor linkage of deafness to markers in the 7 cM region showed a score of 13.55 at markers D5S2119 and D5S2010 (Leon et al. 1992). This was not only the first AD critical region described, but the first autosomal non-syndromic hearing loss gene to be mapped altogether. It was not until 1997,26 years afterfirstbeingreported,thattheregionwas narrowed down further.

Positionalcloning,sometimesreferredtoasreversegenetics,isthecloning of an area known to be associated with a disease. It involves the isolation of overlapping DNA segments that progress along the chromosome toward a candidate gene. Lynch etaI.

(1997)perfonned fine mapping using positional cloning techniques to narrow the critical region to 1 cM. This revealed protein diaphanous homolog 1(DIAPHI), a previously unidentified human gene.DIAPHI, in this case, is a positional candidate gene, a gene identified based upon a detennined critical region. Thisdiffersfrom functional candidate genes, which are known to play a role in the disease pathology, or have been previously shown to harbor mutations that cause a disease,like the p.A716T mutation in WFSIin Newfoundland (Young etaI.2001; Bespalova etaI.2001). Lynch et al. (1997) sequenced the positional candidate geneDIAPHI in all affected family members using Single Strand Confonnation Polymorphism (SSCP) analysis. SSCP is the electrophoretic separation of single-strandednucleicacidsbasedondifferencesinsequenceswhichbringabouta

different secondary structure and thus a measureable difference in mobility through a gel.

The causative mutation, a G7T substitution, was now revealed, andDlAPHlwas also found to be higWyexpressed in the cochlea (Lynch etaI.1997). Lynch eta1.(1997) speculate that the protein this gene encodes, protein diaphanous homolog I, plays a role in the regulation of actin polymerization in the hair cells of the cocWea.

Founder Populations& Mutations

No surnmary of hearing loss associated genes and mutations would becomplete without mentioning the irnportanceoffounderpopulations, thefoundereffect,and foundermutations.Afounderpopulationisasmallsubpopulationthat has been isolated due to geography, culture, religion or a combination of these. Thissubpopulationhasa significantly decreased amount of genetic diversity, causing certain genetic traits to either vanish or become very abundant in further generations. When a founder populationis isolated individuals in later generations are likely to share many genes,becausea mutation in a founder will be passed onto a large proportion of the populationin subsequent generations (Nurhousen, 2000). Founder populations, therefore, possess much promiseindetermining the genes involved in genetic diseases. As there is little genetic heterogeneity, the majority of the individuals with a given disease will carry the same gene mutation. For example, the Ashkenazi Jews were a reproductively isolated population in Europe for rougWy a thousand years, with very littleout-migration or inter- marriage with other groups (Nebel etaI.2005). As a result of this event, the GJB2

mutation c.167delTwas found to be highly prevalent in the Ashkanazi Jewish population (Morell et aI. 1998). Other examples of founder populations include the Canadian province of Quebec, which was established by as few as 2600 individuals, the United States Amish population, and the population ofPingelap, a small island in Micronesia.

A founder mutation is a mutation found as an allele and shared by several individualsfromafounderpopulationandderivedfromasingleancestor.Forexample, CaCHp.P51P/S mutation carriers are considered to have originated from a common ancestor (de Kok et aI. 1999). A second example was recently seen when Park et aI.

(2010) investigated the 3-bp deletion in intron 7 (c.991-15_991-13del) ofDFNA5, and identified a conserved haplotype between a Korean family and a Chinese family segregating the deletion in DFNA5,suggestingthatthis deletion represented a founder mutation originating from a common ancestor.

Colonizationof Newfoundland: A Founder Population

The island of Newfoundland makes up a large part of the Canadian province of Newfoundland&Labrador.Itis the most easterly landmass on the North American continent.In1497 the European explorer Giovanni Gabotto (John Cabot) "discovered"

Newfoundland, though Vikings had landed earlier. Europeans voyaged across the North AtlanticfromEngland,Scotland,lreland,FranceandPortugaltoharvesttherichfish stocks. When each fishing season ended they returned to their countries, as permanent

those consideringpennanent settlement, and with a lack of basic supplieseven aone year stay would have been very difficult (Poole & Cuff, 1994). This deterrence lessened throughouttheI7^thcentury,however,assmailgroupsofEnglish,Scottish,and Irish settlerssetsailfromwesternEngiandinl6l0andthroughoutthel7'h century.These colonists excluded other nations from fishing off of Newfoundland's east coast, but were discouraged in their settlement by the English government, who saw their presence as a threat to the monopoly control that Western England fishing centers had established.

Meanwhile, fishennan from France dominated the island's south coast and northern peninsula(Bennett,2002). Throughout the l600s, the French began to pennanently settle, but in l713,withtheTreatyofUtrecht, the English gained control of the south and north shores of the island. Pennanent settlement increased rapidly by the latel8^thcentury, peakingintheearlyyearsofthel800s.

The colonization of Newfoundland beganinearnest in the early 19thcentury mainly from Southwest England and Southeast Ireland. Fishennan brought their families totheisland,intendingtosettlepennanently,andthesefamiliesarethefoundersofmuch oftoday's Newfoundland population. Ninety percent of Newfoundland's population descends fromrougWy 30 000 founders (Parfrey etal. 2009). Farniliessettled in small inlets along Newfoundland's coast in groups of one or two families. These communities developed in geographical and cultural isolation, and can be characterized by large families, and a strong founder effect (Bear et al. 1987). This isolation also directly led to manygenerationsofinterbreeding(poole&Cuff,1994;Hancock,1989).

families, and a strong founder effect (Bear eta1. 1987). This isolation also directly led to many generations of interbreeding (poole&Cuff, 1994; Hancock, 1989).

Large extended pedigrees from genetic isolates have been instrurnental in the identificationofgeneticcausesofhereditarydisorders.Severalfounder mutations have been identified in Newfoundland. For example, an exon 8 deletion in MSH2, a gene causing hereditary non-polyposis colorectal cancer, has been found in five different Newfoundland families (N=74 carriers), and an intron 5 splice site mutation (c.942+3A>T) has been found in 12 different Newfoundland families (N=151 carriers) (Frogattetai, 1996; Stucklessetal. 2006). A third example is a founder mutation (c.782+3delGAG) found in the deafness gene TMPRSS3 in two different Newfoundland families (Ahmed eta1. 2004; Young et a1. unpublished data). These are just a few examples of founder mutations identified in Newfoundland, and serve to highlight the importance of genetically isolated Newfoundland families in the study of hearing loss. At the beginning of this study, only six hearing loss associated genes had beenidentifiedin the Newfoundland population (Table 1.2). There is a possibility that a founder mutation exists and could be found in some of these 28 Newfoundland families. When a founder mutation is identified, the prevalence of this mutation in different worldwidepopulations canbecompared,andbetterestimatesofriskforindividualsinthefounderpopulation can be calculated. The presence of founder mutations in Newfoundland could thus have strong clinical irnplications in terms of improved diagnosis and the ability to routinely screen individuals if a founder mutation is common enough in the population.

with Bardet-Biedl syndrome (Webb et al. 2009). So while many Newfoundlanders can trace their roots to roughly 30 000 founders, it is critical to keep inmind that these founders came from different towns and different regions. The current population of Newfoundland&Labrador, according to a 2006 census, is 505 469. A map of Newfoundland&Labrador is seen in Figure 1.3.

A)

:: 8:'f---+--t----+--t---:=:=.-'=---j

rf:=:=+==:::::~~::::::~~;=1

~: f---+--t---t--I----+---j

:: 8::

;

~

~: ^,.--f--_

Figure 1.1 Examples of Audiograms. A) Audiogram of an individual witb normal bearing in botbears. B) Audiogram of an individual witb moderate bearing loss in botb ears. C) Audiogram of an individual with severe bearing loss in both ears.

A)

~ "f---.,..--+---~~----!--l---+---1

! : f---"'---+---+~--'il~~--:I~==--'~l---+---1

~⁷⁰f---'--+---+--*""""F--->~,,~=---+I__..----1

"&--~

Figure 1.2 Examples of More Complex Audiograms A) Audiogram of an individual with moderate hearing loss in the low frequencies sloping upwards to mild hearing loss in the mid-to high-frequencies for the left ear only. Hearingin the right ear is normal. B) Audiogram of an individual with progressive hearing loss showing a mild loss at low frequencies which slopes downwards to a severe bilateral loss in the mid-to high-frequencies.

Figure 1.3 Map of the Island of Newfoundland. AD Families under investigationinthis study are indicated with their geographic location.

Table 1.1 AD Non-Syndromic Deafness Genes Identified Worldwide to Date (Van Camp G, SmithRJH. Hereditary Hearing Loss Homepage.

http://bereditaryhearingloss.org,May2010)

Mu-crystallinhomolo Ion Homeostasis Protein diaphanous homologHairBundle,Cytoskeletal

I Formation

Ga "unctionbeta-3 rotein Ion Homeostasis

KCN04

Potassium voltage-gated channel subfamily KQT

member 4 Ion Homeostasis

Unknown

COCH COLLliA2

Non-syndromichearing im airment rotein5 Unknown

Wolfr Ion Homeostasis

AIha-tectorin Extracellular Matrix

CocWin Extracellular Matrix

E esabsenthomolo 4 Transcription Factor Colla en, type XI,aIha2 Extracellular Matrix POUdomain,class4,

transcritionfactor3 Myosin, heavy chain9, non muscle Actin, ammal M osin-6 Grainyhead-like2 Mosin-Ia Ga "unctionbeta-6 rotein Transmembrane channel- like roteinl Coiled-coil domain- containin~Protein 50

Transcription Factor Hair Bundle, Motor Protein HairBundle,Cytoskeletal Formation Hair Bundle, Motor Protein Transcription Factor

Hair Bundle, Motor Protein Unknown HairBundIe,CytoskeletaI Formation

Table 1.2 Deafness Genes Identified in the Newfoundland Population at Beginning of This Study.

f--~"'---t---"-'-"-=--+-"--"-'--"=="'--ILi~;~~i~~;f:;:~ce

DelCastilloetal.

GJB6 01381830 2002

f--~"""~~~:~~~;---f-c-::;;;.7~~i2~3;;:c:7d~e~I~A=G-+---';:---/ ~~:~::^:::;~~:

f--.:....:;;p~;;;;;~~;-;-:::5-1---"~:;;-;-~i~;.;.;~~;':-';>~'--t---;----t =ede:ta~l}oOoOI~

Chapter 2: Methods&Materials

Human Subjects & Pedigrees

This study is one partofa larger study to detennine the genetic cause of hearing loss in Newfoundland&Labrador. Family members were recruited through the Newfoundland Provincial Genetics Program, and a province-wide ascertainment drive.

Informed consent was obtained from all participants, granting researchers permission to access medical records and family history. Blood samples were collected and genomic and mitochondrial DNA was extracted from peripheral leukocytes from participants.

Audiological tests were performed to determine the type of hearing lossofeachsubject and to confirm normal hearing in unaffected subjects. Audiograms are available for all individuals marked with an asterisk in Figure 2.1, and for each of these individuals several audiograms are available at different test ages, withnewonesroutinelybeing collected. DNA from several Dutch individuals was provided by Dr. Hannie Kremer of the Radboud University Medical Centre in Nijmegen, Netherlands. This project was approved by The Human Investigations Committee (HIC) (Research Ethics Board of Memorial University, Newfoundland&Labrador) (# 01.186).

So far, 128 probands have been recruited to the study. All probands in the study were routinely screened for mutations previously identified to cause hearing loss in the Newfoundland population. Of these, 28 probands are members of multiplex families with a family history of hearing loss consistent with AD inheritance, and were chosen for this

study (Figure 2.1). Inheritance patterns weredeterrnined through an extensivefamily history questionnaire, and pedigrees were electronically stored usingthecomputer program Progeny. Many individuals' hearing loss was determined by word of mouth from family members. In these cases, the age of onset and the degree and severity of hearing loss are not known. For this reason, the determination of an AD pattern of inheritance is not certain in some cases, but these families were deemed to likely have an AD form of hearing loss, and it was therefore worth testing them for potential AD hearing loss mutations. Due to the extent of genealogy work possible in Newfoundland up to this point, it is important to keep other potential formsofinheritedhearingloss, such as mitochondrial inheritance, in mind when searching forcausati vemutations.

Experimental Design: Functional Candidate Gene Mutation Screening

Genomic DNA from pro bands (n=28) was screened using a functional candidate gene approach. A comprehensive literature search was done to collect information on all AD hearing loss genes. One recent review (Hilgert et al. 2009) discussed in depth the genes causing AD hearing loss.It describes how each gene associated with AD hearing loss functions in the ear, and what types of hearing loss they cause. For each of these genes, the mutations found both worldwide and within Caucasianpopulationsare described in detail. Many of these genes may be causative in Newfoundland families (Table 2.1).

This literature search formed the basic foundation from which genotype- phenotype evaluation was performed. Potential candidate genes were investigated for specificcase-by-case details on the hearing loss phenotype each mutationcaused, and in what population and ethnicity they were reported. These phenotypes were then cross checkedwiththephenotypesofthe28farniliestofurthernarrowdownthefunctional candidate gene list to four: COCH, KCNQ4, TECTA, and MYOIA (see Figure 2.2). This approach is a form of audioprofiling, a method of categorizing phenotypic data to make genotypic correlations. The audiological data of several membersinafarnily,orinthis case several probands from different families, associates with a specific unknown genotype as a function of time (Meyer et al. 2007). From this, we have drawn correlations to the overall phenotype of the group of probands and used this as afoundation for selecting candidate genes previously reported to cause hearing loss with a similar phenotype.

Information on all known hearing loss mutations in these four candidate genes was next collated, including which domain and exon each mutation was reported in (Appendix A). This allowed the identification of the exons most likely to harbor causative mutations in the Newfoundland probands. For example, the majority of mutations reported in COCH are reported in the factor c homologous (FeR) domain, spanning exons 4 and 5, and so this area of COCH was screened first.

COCHhas a total ofl2 exons, and causes a late-onset, progressive hearingloss most often associated with vestibular dysfunction (Kemperman et al. 2005), which correlates with several AD Newfoundland families. COCH is important in maintaining

structural support within the cochlea (Kommareddi etal. 2007). Exons screened were 2-5, and 12. Two deletions within KCNQ4 have been reported to cause a late-onset, progressivehearingloss(CouckeetaI.1999;Kamadaetal.2006),matchingthe phenotype of many AD Newfoundland probands. This gene is critical in ion homeostasis with the ear (Kubisch et al. 1999), and is coded by 14 exons, all of which werescreened.

Missense mutations withinTECTA,a second gene important in structural support within the auditory system, have been shown to cause late onset hearing loss 0'erhoevenetal.

1998). Coded by 23 exonsintotal, exons 5, 9-14, 17, 18, and 20 were screened. The last candidategene,MYOIAisthoughttoplayaroleinsoundprocessingthroughion transport (Donaudy et al. 2003), and again, causes alate-onset, progressive hearing loss phenotype. Coded by a total of28 exons,exons3,4,6,7, 10-12, 18, and 22 were screened.

Whilethemore"targeted"candidategeneapproachoutlinedaboveislikely to lead to the identification of mutations, it does restrict the chance of identifying potential

"genetic surprises" regarding genotype-phenotype. This method allows for a complete investigation of possible genetic mutations in a given set of promisingcandidategenes within the two-year time frame of aMaster's thesis. However, exons within these candidate genes that have never before been associated with hearingIossmutationsare bi-directionallysequencedinthisstudy,andsothereispotentialfor"genetic surprises".

Because Newfoundland is a founder population, it is possible that families would share the same mutation, particularly if those families are from the same geographicarea.

When a mutation is discovered in an AD Newfoundland proband, all otosclerosis, AD,

and AR Newfoundland probands (n=68) are subsequently screened for that mutation.

When appropriate, apparent founder mutations are confirmed by haplotype analysis, to determine the level of sharing fora selection of linked microsatellite markersbetween families with the same mutation (see Figure 2.2).

General Strategy for PCR and Sequencing of Candidate Genes

Both forward and reverse strands of specific exon PCRproducts in each gene were bi- directionallysequenced,alongwithallintron/exonboundariestoensurethe entire coding region of each exon was covered. PCR primers were designed using Primer 3 software (v.

OA.0,http://frodo.wi.mit.edu/primer3/).Theseprimersequencescanbe found in AppendixB.

DNA Preparation, PCR Thermocycling, and Electrophoresis

DNA was extracted from whole blood and diluted tolO nglill. This blood was stored at 4°c(performed by research assistant). I ilL of diluted (stock) DNA was added to 2 ilL lOX PCR Buffer (containing MgCh), OAIlL dNTPs (10 mM), 0.08 ilL KapaTaq DNA Polymerase (5 UlIlL) (Kapa Biosystems, Boston, MA), 12.92 ilL of distilled dH20, 1.0 ilL offorward primer (10 11M) and 1.0 ilL of reverse primer (10 11M), as per standard PCR protocol. The amount of dH20was reduced to add betaine or Dimethyl Sulfoxide (DMSO) when necessary to achieve a successful amplified PCR product. This mix was

centrifuged and addeqd to wells in 20 flL aliquots in a 96-well PCR plate, where it was then sealed, centrifuged, and placed in the GeneAmp PCR System 9700 thermal cycler (Applied Biosystems, Foster City, CA). PCR products were electrophoresed on a 1%

agarose gel (1.5 g agarose/lOO mL TBE) stained with SybrSafe and viewed under UV light on a Kodiak GEL LOGIC 100 Molecular Imaging system (Rochester, Y, Version 4.01,2005).

Preparation for ABI Cycle Sequencing

Sephacryl S-300HR was resuspended and 300 flL a1iquots were added to wells on a Millipore Multi-screen HTS plate, which was placed over a corresponding 96-well waste plate to catch flow-through. Plates were then balanced and centrifugedat3000rpm for 5 minutes. Flow-through was discarded and PCR products were added to wells on the Multi-screen HTS plate. The Multi-screen HTS plate was then positioned and placed over a clean PCR plate, balanced, and centrifuged at 3000 rpm for 5 minutes. The flow- through product collected in the PCR plate contains the purified PCRproducts.

Successfully amplified PCR products, visualized as bands of the correctsize (using a 100 bpmarker) when electrophoresed on an agarose gel, were cyclesequenced using the following reaction mix: 0.5 flL of Big Dye Terminator V. 3.1 Sequencing Mix, 2 flL of 5X sequencing buffer, 0.32 flL of Primer (10 flM), I flL of purified PCR product DNA template, and 16.18 flL of dH20, per the Big Dye Terminator V. 3.1 protocol (Applied Biosystems, Foster City, CA). This equals a total reaction volume of20 flL per

well in a sequencing plate. The resulting plate was centrifuged briefly, loaded onto the thennal cycler, and subjected to a thermal cycling program according to ABI BDT V. 3.1 protocol (Applied Biosystems, Foster City, CA).

Upon completion,S ilL of 125mMEDTA followed by 65 ilL of95% Ethanol (EtOH) was added to each reaction well. Plates were briefly centrifuged and then incubated overnight in the dark at ambient temperature. The plate was then centrifuged at 3000xgfor30minutes,invertedtodecantEtOH,andbrieflycentrifugedwhiIe inverted at 200 rpm for4-5 seconds with folded paper towels placed undemeaththesequencing plate to absorb residual ethanol. ISO ilL ono % EtOH was added to each sample, and the plate was centrifuged at 3000 gfor 15 minutes. The plate was again invertedtodecant ethanol and spun at 200 rpm for 4 - 5 seconds over a paper towel. Samples were left to air dry in the dark at room temperature for 10 - IS minutes. IS ilL of Hi-Di Formamide was subsequentJy added to each well and the plate wasvortexed and centrifugedbriefly.

The final mix was denatured at 95°C for 2 minutes on a thermal cycler. Once denatured, samples were kept on ice until placed in the ABI 3130 XL DNA Analyzer.

Automated Sequencing Using the ABI 3130 XL

Automated sequencing was performed using either the ABI 3130 XL DNA Analyzer (Applied Biosystems, Foster City, CA) availableinthe lab or the ABI 3730 DNA Analyzer in the Genomic & Proteomics Facility, at CREAlT, Memorial University of Newfoundland. The raw sequence data were initially analyzed for quality using